The author touches on this in the last section, but I'd reframe this a different way. The natural conclusion for a company who wants to funnel you to the app is, "the web version is a-OK? Let's make the web version worse."

I'd rather see this framed as, "if you don't have a high functioning web version, I don't need to use your service." Gimping my preferred medium will lose me as a customer. If enough people draw that line, "enshittifying" your web app should hurt your metrics, not help. That way maintaining a good web version is looked at as a long-term necessity, not a top of funnel.